Many Pasts

The description states that Many Pasts is an AI-assisted visual history platform. It allows users to ask historical questions or provide narratives, which are then converted into structured visual programs using GPT-5.6. These programs include narration, subtitles, dates, camera framing, map actions, routes, environmental layers, images, diagrams, evidence, and uncertainty labels.

It includes two main interfaces:

• A Viewer for learning and exploration.
• A Director workspace for editing shots, narration, visual cues, media, and sound before producing a video.

The system is built using technologies such as Next.js, React, TypeScript, MapLibre, PostgreSQL, PostGIS, and the OpenAI Responses API. It features:

• A historical ontology and evidence-aware knowledge graph
• Temporal map layers for geography, climate, ice, hydrology, civilizations, resources, and movement
• A shared frame clock synchronizing maps, narration, subtitles, media, and camera movement
• Local-first spatial database and data-import pipeline
• Narration, captions, audio editing, and deterministic video rendering

Not evidenced: No information on actual user base, usage metrics, or production deployment details.

The author claims that Many Pasts aims to help visual learners understand history by showing how events connect across time, place, people, and ideas. It positions itself as an alternative to traditional timelines (which reduce history to a sequence of dates) and static maps (which show only one frozen moment).

It also states that the platform supports multiple forms of content creation:

• Lessons
• Questions
• Documentaries

The project evolved from an idea carried for a long time into a working prototype during OpenAI Build Week. The author notes that GPT-5.6 powers lesson-planning workflows, but emphasizes that control over evidence, geometry, timing, safety, and presentation remains with the application.

Inference: The positioning suggests a niche audience focused on education, research, and storytelling in history. However, no claims about market size or competitive differentiation are made beyond self-description.

The description states that Many Pasts is intended for:

• Students
• Educators
• Visual learners
• History enthusiasts
• Content creators

It also mentions specific use cases such as:

• Exploring historical narratives
• Creating educational videos
• Inspecting evidence behind stories

Not evidenced: No data on actual users, customer segments, or buyer personas. There is no indication of how many people are using the product or what their behavior looks like.

The description does not contain any information about pricing models, monetization strategies, or business model assumptions. The author focuses entirely on the technical and creative aspects of building the platform.

Inference: Since this is a hackathon submission with no mention of revenue, customers, or sales, it appears to be an experimental project without a defined commercial path at this time.

The system uses:

• Next.js
• React
• TypeScript
• MapLibre
• PostgreSQL
• PostGIS
• OpenAI Responses API
• Docker and Docker Compose
• Faster Whisper for audio processing
• Playwright for testing
• HTML5, WebGL, GeoJSON, GeoTIFF

Key technical features include:

• Historical ontology and evidence-aware knowledge graph
• Temporal map layers with changing environments
• Strict structured format for AI-generated visual lessons
• Shared frame clock for synchronization
• Viewer and Director interfaces
• Local-first spatial database and data-import pipeline
• Deterministic video rendering

Not evidenced: No details on scalability, infrastructure, or deployment practices beyond the prototype.

The project is described as a prototype built in three days during a hackathon. It includes:

• A flagship Mesopotamia experience
• Examples covering global deep-time, Doggerland, palaeocoastline, climate, tectonic, mythology, linguistics, trade, and military topics

However, there is no evidence of:

• Users or customers
• Revenue or monetization
• Product adoption or retention
• Market feedback or iteration history

Inference: The project is at a very early stage—likely pre-product-market fit—and lacks any measurable traction.

The description does not reference existing competitors or market positioning. It does not state whether similar platforms exist, nor how Many Pasts differentiates itself from them.

Not evidenced: No competitive analysis, benchmarking, or market landscape provided.

• Single-person development: The team size is listed as one (Dax Riven), raising concerns about scalability and long-term maintenance.
• Prototype-only status: The entire product appears to be a hackathon prototype with no indication of production readiness or user engagement.
• Unclear commercial viability: No pricing, monetization, or customer base described.
• AI dependency without control: While GPT-5.6 is used for planning, the system maintains strict control over validation and presentation—this may limit its ability to scale without further development.
• Complexity of historical data integration: Handling datasets with different projections, resolutions, timescales, and evidence standards presents a significant challenge.

1. What is the current status of the product? Is it being used by educators or students?
2. How do you plan to monetize the platform?
3. Have you validated demand for this type of tool among your target users?
4. What are the key challenges in scaling the historical knowledge graph and data integration?
5. How does the system handle conflicting interpretations of history?
6. Are there any partnerships or institutional collaborations in progress?
7. What is the roadmap beyond the current prototype?